一种盲信号处理中的时域自适应滤波算法

摘要

滤波是信号处理中的重要环节,鉴于盲信号处理本身的特点,传统的滤波技术并不适合直接用于盲源分离之中.然而作为分离前的预处理,滤波技术在独立成分分离算法中是必要的.为此,结合稳健的数据非线性投影,首次提出盲信号中的自适应滤波方法,与此同时给出了具有自适应特性的阈值判决.在此基础上构造了盲信号中的自适应滤波算法,解决了利用低通和高通滤波处理盲信号所遇到的问题.仿真结果表明,在不破坏数据统计特性的前提下,该方法能有效滤除数据中的野值成分,避免了野值数据对独立成分分离算法性能的影响,为盲信号分离的预处理开辟了一种新的途径.%Filtering is an important part of signal processing.In view of the characteristics of blind signal processing(BSS),the traditional filtering technology is not suitable for blind source separation directly.However,as a pre-processing technique,filtering is necessary to use in the separation of independent components.This paper proposed an adaptive filtering method for BSS.Meanwhile,it also gave a judgment method for the threshold of filtering.Then,it proposed the algorithm of adaptive filtering for BSS.With this method,it could solve the problem arise by using low pass and high pass filtering for blind signal processing.The simulation results show that this method not only can find out the outliers components in the observed data effectively,but also do not destroy the statistical characteristics of the data.Further,it can avoid the poor performance that the independent component algorithm use higher order statistics.